AI's Inevitable Photographic Journey
Head's up that this post is still a draft. It's mostly done. Just gotta figure out how to stick the landing and add footnotes.
The AI hype is profound. I've written a few pieces to help clarify my thinking about it. This one considers the parallels between AI and the history photography.
The Origins
Photography used to be hard ass work.
After folks figured out there are chemicals that turning darker when exposed to light some mad genius realized the images cast in a camera obscura came from light and they put the chemicals on the projection surface and photograph was bornphotohistory.
Took some doing though. You can to build your own camera obscura. You had to get your hands on the chemicals. You had to mix them without killing yourself. You had to coat the emulsion on the paper yourself and run the exposed sheet through more chemcial baths to finish the development process.
Every camera was a hand built contaption of it's own design.
Computers used to be hard ass work.
After folks figured out you can hook up some of the components used by telegraph and phone companies to make littl machines that counted up every time you hit a button some mad genius realized you could wire thousands of them together, use the numbers to represent letters or colors or anything and computers were borncomphistory.
Took some doing though. You had to get access to the components. You had to wire them yourself without killing yourself. You had to input every single number manually by clicking a button and flipping switches.
Every computer was a hand built contraption of its own design.
For the (Rich) Masses
Years pass. Companies saw potential and started mass producing cameras.
Only rich folks could afford them, but it was no longer necessary to build your own. A new way to create was introduced to the public.
Years pass. Companies saw potential and started mass producing computers.
Only rich folks could afford them, but it was no longer necessary to build your own. A new tool to create was introduced to the public.
Cheaper and Cheaper
Technology is refined. Quality improves. At the same time, the cost of production drops.
More folks can afford to buy the cheaper cameras. Every year they get better and better. But, the fundamentals of the box and the film inside it stay the same.
Technology is refined. Quality improves. At the same time, the cost of production drops.
More folks can afford to buy the cheaper computers. Every year they get better and better. But, the fundamentals of the box and apps they run stay the same.
Technology Advances
Digital Sensors are invented. They're branded as DSLRs.
They provide a new way to produce photographs. The box is the same. They still caputre light and produce images, but what goes on inside the box happens in a differnet way.
They suck at first. Only researches use them. But, every generation is bounds better than the last.
When they're introduced to the public they still suck. They images only look good in certain cirsumstances. They fail completely in others. But, they imporve every time a new model is released.
Old school photographers turn up their noses at them.
Large Language Models are invented. It' branded as AI.
They provide a new way to interact with computers. The box is the same. They still take input and provide output, but what goes on inside the box happens in a different way.
They suck at first. Only researches use them. But, every generation is bounds better than the last.
When they're introduced to the public they still suck. The answers are only accurate in certain cirsumstances. They failed completely in others. But, they imporve every time a new model is released.
Old school computers users turn up their noses at them.
Brand New Access
Companies start putting sensors in new places. Phones, specifically.
It's new and novel at first. Limited to specific phones. But, they proliforate quickly. Soon it's virtually impossible to get a phone without a camera in it.
Cameras go from independent devices to merely being a part of phones. Folks who never would have otherwise bought a digital camera end up with one in their pocked by default.
The marketing of phones becomes more about the camera than the phone itself.
Old school photographs talk about how crappy the images are and how digital will never be as good as film.
Companies start putting AI in new places. Everywhere, generally.
It's new and novel at first. Limited to specific apps. But, it proliforates quickly. Soon it's virtually impossible to get an app without a AI in it.
AI goes from independent apps to merely being a part of other app. Folks who never would have otherwise bought used AI end up with it in their apps by default.
The marketing of apps becomes more about the AI than the app itself.
Old school techies talk about how crappy the answers are and how AI will never be good.
You Are Here
Improving Iterations
Technology is refined. Quality improves while costs drop.
Images taken in different types of light no longer look like shit. There's still a bunch of times when they fall apart though.
Over time, those problems become fewer and farther between.
Technology is refined. Quality improves while costs drop.
Responses of different types are no longer wildly inaccurate. There's still a bunch of times when they fall apart though.
Over time, those problems become fewer and farther between.
The New Normal
The general public no longer thinks cameras in phones are remarkable. It's just the way things are. Company marketing still tries to make a huge deal out of upgrades but largely no one cares. The quality is pas ta threshold that's good enough for most folks that they stop thinking about it.
Image quality is mostly good, most of the time.
All but the most ardent hold-outs shift to using the tool. The quality isn't flawless though. Every iteration gets better, but there are always cases where adjstments are required.
The general public no longer thinks AI in apps is remarkable. It's just the way things are. Company marketing still tries to make a huge deal out of upgrades but largely no one cares. The quality is past a threshold that's good enough for most folks that they stop thinking about it.
AI results quality is mostly good, most of the time.
All but the most ardent hold-outs shift to using the tool. The quality isn't flawless though. Every iteration gets better, but there are always cases where adjstments are required.
The Camera and the Violin
There's an old photography riff that goes:
If you have a camera, you can make a beautiful picture.
If you have a violin, you have a violin.
The idea has two parts:
- Cameras are so good at what they do that folks can make a beautiful image the first time they pick one up, even if it's accidental.
- Playing a violin has a much higher barrier to entry. You can't accidentally play something beautiful. It takes a lot of practice to produce anything other than noise.
In the earliest days of photography, that wasn't true. Without the knowledge of how to combine the chemicals to make light sensitive emulsions, you couldn't make a picture. But, technology advanced. Everything that made the statement not true was whittled away until it was.
In fact, we've gone farther. The adage is from the days before cameras were baked into our mobile devices. Updating it for today:
If you have a phone, you can make a beautiful picture.
If you have a violin, you still just have a violin.
The history of computers has been a march from violin to camera to phone. At first, only trained practitioners could do anything useful. As technology advanced, more people could do more things. An expansion of possibilities accompanied by a flattening learning curve.
AI is a Cambrian explosion. The barrier to entry is nothing more than access to the internet. A few words dropped into a prompt can generate reports, books, photos, movies, apps, whatever...
The Eye of the Beholder
Most photos aren't beautiful. The mechanics and software of the cameras built into our phones ensures they're properly exposed. That they're good from a technical perspective. Those mechanics can't guarantee beauty.
But, lots of photography isn't about beauty. It's about emotion. A reminder of how you felt. That's why folks aren't interested in seeing your vacations photos. You're remembering the great time you had at dinner. All they see is the dark exterior of a restaurant called "good sushi".
Footnotes
The history of photography is more complicated. That's an entirely other post though. The minutia of the origin isn't the point. The fact that it took a lot of work to make an image is what matters for our purposes.
Again, the history of computers is more complicated. That's an entirely other post though. The minutia of the origin isn't the point. The fact that it took a lot of work to make a program is what matters for our purposes.